Incidence of probable occupational asthma and changes in airway calibre and responsiveness in apprentice welders

Since the 1980s, epidemiological studies investigated the effects of exposure to welding fumes and gases on lung function and clinical manifestations in welders. The assessment of the effect on pulmonary function is limited, in that most studies have been cross-sectional in design, conducted ingroups of welders and nonwelder referents. These studies have not consistently demonstrated an impairment of pulmonary function in occupationally exposed welders 1. However, there is now good evidence of a higher prevalence of ventilatory impairment in welders shown by significantly lower pulmonary function indices than in comparable controls 2–9. The confounding effect of smoking has been demonstrated 5, 6, 10. Studies of short-term respiratory function have shown a greater reduction in lung function indices among welders than among nonwelders 3, 11.

There have been relatively few published longitudinal studies. A 2‐yr follow-up study showed no significant overall difference in the annual change in pulmonary function variables between welders and nonwelders; however, smoking welders and those welding without local exhaust ventilation or respiratory protection had significantly increased risks for accelerated decline in the forced expiratory volume in one second (FEV₁) 12. This confirmed earlier findings by Chinn et al. 13 in a group of shipyard welders. A 3‐yr prospective cohort study on airway reactivity in a group of shipyard workers, including 51 arc welders and 54 nonwelders, showed that welding was associated with a transient across-work-shift decrement in midflow and reversible work-related symptoms; however, no chronic irreversible effects on spirometry or bronchial responsiveness was seen over the 3 yrs 14.

Bronchial responsiveness to specific agents from the working environment and to nonspecific (methacholine or histamine) agents has not been extensively investigated in welders. A cross-sectional study on stainless-steel (SS) and mild-steel (MS) welders found that bronchial responsiveness and lung function in active welders was normal and did not differ between MS and SS welders or between welders and a reference group of vehicle assembly workers who had never welded 8.

To the best of the authors' knowledge, no prospective studyof the incidence of bronchial hyperresponsiveness (BHR) has been performed in a population of welders with objective measures of bronchial responsiveness at the onset ofexposure. Moreover, given the inconsistency in the previously conducted studies, further investigation on changes in bronchial responsiveness and airway obstruction, especially in asymptomatic subjects, is needed in welders starting exposure to metal fumes.

The objectives of the study were to determine, in apprentice welders: 1) the incidence of welding-related respiratory symptoms coexisting with significant BHR; 2) the incidence ofBHR; and 3) the decrease in FEV₁ brought about by exposure to welding fumes and gases for an average duration of 15 months.

Methods

Results

This group of apprentice welders represents a relatively young population with a mean±sd age of 24.5±7.4 yrs, dominated by males (85.7%). Slightly more than one-third (33.7%) had never smoked, and 50.5% were current smokers.

Table 1⇓ shows the incidence of respiratory symptoms at thetwo follow-up periods as well as the results of the lung function and methacholine bronchial challenge tests at baseline and at the end of study. As shown, data are presented for 194 apprentice welders. The incidence of respiratory symptoms was similar at the first and second follow-up assessments. A statistically significant difference was found between the baseline and end of study mean values for the lung function tests. In particular, FEV₁, % pred had significantly dropped by 8.4% on average. A higher percentage of subjects had a FEV₁, % pred <80% at the end of the study compared with the baseline values (11.3 versus 3.6%, p<0.01). A statistically significant intra-subject difference was found in the slope of the methacholine dose/response curve; the change indicated an increase in bronchial responsiveness (i.e. higher value at the end of the study). Nearly 14% of the 194 apprentices had a PC₂₀ ≤8 mg·mL⁻¹, and 20% had a PC₂₀ ≤16 mg·mL⁻¹ at baseline. These figures were higher, nearly 19 and 24%, respectively, at the end of the study. The proportion of subjects with a PC₂₀ ≤8 mg·mL⁻¹ at follow-up was significantly greater than at baseline.

Table 2⇓ shows that a greater proportion of subjects (15.2%) with baseline PC₂₀ ≤16 mg·mL⁻¹ compared to subjects with baseline PC₂₀ >16 mg·mL⁻¹ (6.5%) had a drop in FEV₁, %pred <80%, although this was not significant (p=0.15). Having a baseline PC₂₀ ≤8 mg·mL⁻¹ (20%) compared to a PC₂₀ >8 mg·mL⁻¹ (5.7%) was associated with a decline in FEV₁ % pred <80% (p=0.02) (results not shown). Then again, having a steeper slope of the methacholine dose/response curve at baseline (1.72 versus 1.53) was not associated with a drop in FEV₁ % pred <80% (p=0.19) (not shown).

Conversely, when considering the decline in FEV₁ in termsof mean % decline, no difference between subjects according to PC₂₀ at baseline whether comparing subjects with PC₂₀ ≤16 mg·mL⁻¹ to those with PC₂₀ >16 mg·mL⁻¹ (−7.48±11.44% pred versus −8.68± 8.56 % pred, respectively) or when using the cut-off value of PC₂₀ 8 mg·mL⁻¹ was found. In addition, no correlation was found between mean FEV₁ % decline and dose/response slope at baseline.

The pattern of change in bronchial responsiveness during the course of the apprenticeship is detailed in table 3⇓. Nearly 7% had a 3.2‐fold decrease in PC₂₀ when they initially had abaseline PC₂₀ >16 mg·mL⁻¹. On the other hand, slightly over 5% had that decrease while their baseline PC₂₀ was ≤16 mg·mL⁻¹. Thus, a total of 23 subjects (11.9%) had a significant increase in BHR (3.2‐fold decrease in PC₂₀). Incontrast, using a less conservative criterion (a two-fold decrease in PC₂₀), 29 subjects (14.9%) had an increase in BHR. The mean decline in FEV₁ % pred in subjects with a significant decrease in PC₂₀ (>3.2‐fold decrease) was −10.7 compared to −8.13% among the other subjects (p=0.21).

Table 4⇓ illustrates the distribution of subjects with at least one welding-related symptom according to whether or not they showed an increase in bronchial responsiveness. The incidence of welding-related respiratory symptoms suggestive of OA was 13.8% throughout the study with participants experiencing more cough (9.9%) than wheezing or chest-tightness (5.2 and 5.6%, respectively). Seven subjects persistently reported having at least one of the following symptoms: cough, wheezing and chest-tightness (data not shown). Of the 20 subjects who had a ≥3.2‐fold decrease in PC₂₀ from the baseline value, five apprentices had persistent respiratory symptoms suggestive of OA. Thus, the incidence of probable OA was in the order of nearly 3% (6 of 194 or 5 of 194) when there was either a two-fold or a ≥3.2‐fold decrease in PC₂₀ from one visit to the next.

Table 5⇓ presents a detailed description of probable cases ofOA. It shows that three apprentices with normal airway responsiveness at baseline had a significant reduction in their PC₂₀ by the end of the study. The other three apprentices with pre-existing airway hyperresponsiveness had a further deterioration expressed by a ≥3.2‐fold reduction in their PC₂₀. Apprentices mostly reported the presence of welding-related respiratory symptoms suggestive of OA at the second follow-up, particularly symptoms of cough and chest-tightness. Only one apprentice-welder reported having a physician-diagnosis of asthma.

Discussion

To the best of the authors' knowledge, this is the first prospective study to measure the incidence of airway obstruction, BHR and OA in apprentices newly exposed to welding fumes and gases. It was found that: 1) FEV₁ % pred dropped significantly (by 8.4% on average) from the pre-exposure baseline value; 2) 23 apprentices (11.9%) had a significant increase in BHR when it was defined as a ≥3.2‐fold decrease from baseline PC₂₀ and; 3) the incidence of probable OA, using a definition that the authors proposed, was nearly 3%.

The drop in FEV₁ % pred was important but probably cannot be attributed to a chronic effect of exposure to welding fumes and gases, given the relatively short duration of thewelding training programme. The tests undertaken were being performed at the teaching institutions during a normal class day. Although some authors found no significant differences between lung function changes over the day between welders and controls 6, 25–27, others have found significant decreases from morning to afternoon in the mean change in FEV₁ and FVC among both welders and controls, but the reduction was almost four-times greater among welders 3 where in the latter study welders who had the reduction in the pulmonary function indices did not differ from those without reductions in terms of age, height, baseline lung function or smoking habits.

In the present study, spirometry and methacholine tests were performed on two occasions, 15 months apart on average, first before the apprentices started exposure to welding fumes and gases, and again at the end of the vocational training. The effect of independent factors (i.e. day of the week the test was carried out, baseline airway responsiveness and smoking) on the decrease in FEV₁, % pred, other than exposure, was examined. Assuming that a weekly effect would be present, the proportion of subjects with a drop in FEV₁ % pred would have to be greater for those who had a test performed between Tuesday and Friday at the end of the welding training programme. On the contrary, when taking into consideration whether the second methacholine test was done on a Monday or another day of the week, there was stilla significant drop in FEV₁ % pred (9.87±10.15% and 8.20±9.05%, respectively), regardless of the day of the week on which the follow-up test was performed. Upon considering whether having BHR at baseline would be associated with a decrease in FEV₁ % pred value; the results showed very similar changes in mean decline FEV₁ whatever the level of bronchial responsiveness at baseline. With regard to the dropin FEV₁ % pred <80% the proportion of subjects with baseline PC₂₀ ≤16 mg·mL⁻¹ (the present criteria for BHR) was greater than that with PC₂₀ >16 mg·mL⁻¹ suggesting an association, although not significant; the lack of significance could be due to a lack of statistical power, since the numbers of subjects with a drop in FEV₁ % pred <80% was small (n=15). Finally, there was no significant difference in the meandecline in FEV₁ % pred between smokers (8.92±8.97%) and nonsmokers (8.02±9.40%). Deterioration in pulmonary function among welders compared with a reference group 2–9, has been previously noted in epidemiological studies, with greater effects observed in smokers 6, 13, 27, 28, confirming the role of smoking as the main risk factor leading to the decline in lung function. The absence of an influence of smoking in this study could be partly explained by the current study subjects being young adults who consequently would not have been significantly affected by respiratory conditions due to cigarette smoking. Moreover, the same well-maintained instruments and a standardised procedure for performing spirometric and bronchial responsiveness tests were used. Hence, the authors believe that the demonstrated functional impairment of the apprentice welders most probably reflects actual changes in the level of bronchial calibre and that the exposure to welding gases and fumes does appear to decrease pulmonary function parameters acutely in this context.

When comparing the 10 subjects with a baseline PC₂₀ >16 mg·mL⁻¹, who had a drop in their FEV₁ % pred <80%, to the 144 students who did not have that drop it seems that from the start the former had, on average, significantly lower values of FEV₁ % pred (90.93±5.86% and 110.69±12.90%, respectively). By considering these 10 subjects separately, it isfound that one-half of them had a FEV₁ % pred <90%. Three were current smokers who also reported symptoms suggestive of asthma at baseline. It was found that two of these three also became hyperresponsive to the methacholine challenge. One of these two was a smoker who reported having had asthma diagnosed by a doctor, and who had wheezing and chest-tightness in the year prior to the baseline interview as well as at the last follow-up and metal fume fever. However, this subject did not report any welding-related respiratory symptoms.

In the current study, ∼29% of the apprentices were lost to follow-up with the majority having quit the apprenticeship shortly following the baseline assessment survey. Although these were not contacted, it is believed that the short duration between enrolling and dropping out excludes possible self-selection out of the apprenticeship because of welding-related health complaints after entry into the programme. Those who were lost to follow-up in the current study did not differ in terms of baseline demographical characteristics from those who remained until the end of apprenticeship. Close examination of baseline symptomatology revealed that a significantly higher proportion of apprentices lost to follow-up had phlegm, while a significantly lower proportion of apprentices lost to follow-up had a personal atopic history (history of eczema, urticaria, or hay fever), rhinitis with pets and in the pollen season, as well as respiratory symptoms with pets.

An interesting finding of this study was that the incidence of BHR was higher than that of OA. As many as 75% (15 of20) of the apprentices who had a 3.2- fold decrease inPC₂₀ had no persistent welding-related respiratory symptoms suggestive of OA, although eight of 15 subjects reportedrespiratory symptoms at one of the follow-up visits. Increasedbronchial reactivity to nonspecific bronchoconstricting stimuli is a characteristic property of asthma, but BHR does not constitute asthma and is found in some individuals with no history or symptoms of asthma 23, 29. Approximately 50% of subjects with airway hyperresponsiveness report no respiratory symptoms 30, 31. In a 3‐yr follow-up study 32, subjects with airway hyperresponsiveness and no respiratory symptoms had a greater increase in airway responsiveness and incidence of asthma symptoms than symptomatic asthmatic subjects or normoresponsive controls. These authors, along with others 33, suggested that airway hyperresponsiveness appears to be an intermediate stage between normality and symptomatic asthma. More recently, Gautrin et al. 34 in a prospective study of 417 apprentices in animal-health technology have shown that pre-exposure airway calibre and responsiveness are associated with an increased risk of developing probable OA.

In this study, the criteria for defining probable cases of OA included the presence of welding-related respiratory symptoms together with BHR upon considering previous definitions for asthma and OA. Toelle et al. 35 defined current or common asthma as BHR plus recent wheeze (occurring in thelast 12 months prior to the study) for measuring the prevalence of clinically important asthma in populations. OA has been defined as “a disease characterised by variable airflow limitation and/or BHR due to causes and conditions attributable to a particular occupational environment and notto stimuli encountered outside the workplace” 36. Clinically, OA is manifested by work-related symptoms of chest-tightness, wheezing and cough, while physiologically, there are alterations in lung mechanics that change with time 37. In a previous prospective study among apprentices exposed to laboratory animals, the incidence of probable OA was found to be in the order of 2.7% 34. The incidence of OA that was found in apprentice welders is very close to this. In the quoted study, probable OA was defined as the onset ofimmediate skin reactivity to one or more occupational inhalants together with a ≥3.2‐fold decrease in PC₂₀. Moreover, it has been reported recently that the sensitivity of clinical history, nonspecific BHR, and skin-prick tests against natural rubber latex (high molecular-weight agent) in the clinical assessment of OA, are in the order of 87, 90 and 100%, respectively 38.

Conclusion

Overall, the prospective study provided increasing evidence that exposure to welding fumes and gases is associated with pulmonary functional changes and respiratory symptoms in welders. These changes have been seen in the apprentices over an average short period of 15 months. Since apprentices will become newly hired welders and thus part of the future welding population, it is relevant to examine whether these changes represent a predisposing factor to further chronic abnormalities.